Saying something specific: vague, empty and loaded words
A word can be correct and familiar and still say nothing a reader can use, or say more than the writer meant. It can sound like content without saying what was meant, overstate or soften a claim, stand in for a thought the writer has not finished, gesture at a gap in the literature or at novelty, carry a legal reading the writer did not intend, or date a claim that will be read after it was written.
How specific a word is
The reader should not need to interpret what the writer has said. For medical writing, the marketing maxim keep it short and simple can become keep it simple and specific, which asks of each bullet point whether it is specific. Each word should be checked for whether it is as specific as possible, and the check is better made at the outline or draft stage than in the whole article: there it is easier to see, and once the long writing has begun, what is missing is too hard to see. Once a word is specific in the outline, the sense is not lost in the full article. Writers working in a second language should go back and check whether each word describes what really happened, and change it where it does not.
Empty words and sentences
Quality is an empty word, and even good quality or bad quality does not always say what is meant. The writer should try to avoid it and ask which quality is meant: cheaper, more effective, compliant with the guidelines, cost-effective or better for quality of life.
A whole sentence can say nothing, such as this one from a document under editing: "Some different types of leukemia have been found to respond differently to different treatments." It is like Spock telling an evil android he meant to disable "Everything I say is a lie. I am lying", at which the android melted down.
Changes, outcomes and values stated incompletely
Blood pressure changed means nothing when the reader knows neither where it started nor where it finished, so the word should say whether the value increased or decreased. The patient reacted well depends on the patient's state beforehand, so the writer should say more specifically, for example, how much the patient improved. Treated says nothing until it says treated with what. The patient's blood pressure was low is not wrong in itself, but for a patient who always had low blood pressure it is not a problem, so the reader needs to know compared with what.
Words that overstate or soften a risk: safe and potential risk
Many medical writers and translators may have noticed that safe and safety, alone and in compounds such as safety evaluation, safe procedure and safety data, are used increasingly often, particularly in texts pharmaceutical companies write for investigators (protocols and investigator's brochures), for health authorities (the Summary of Product Characteristics) and for trial participants and patients. In the view of two medical translators, both words are euphemisms whose meaning is vague: a reader cannot tell whether a safe drug has no adverse effects or has them only infrequently or only minor ones, and calling a drug safe might suggest that it has none. There is no zero risk in clinical medicine, which means that all diagnostic and therapeutic procedures entail risk, while dictionaries define safe first as being free from, not in, or protected from danger or harm. In the translators' view, such vague words act against clarity and precision, which should always be the principal objective of scientific language, and calling a drug safe when advertising it to consumers is tantamount to deceptive advertising. The same translators hold that the health authorities that regulate medicines, such as the Food and Drug Administration, the European Medicines Agency and the Spanish Agency of Medicines and Medical Devices, should forbid safe and safety in texts addressed to trial subjects, investigators, patients, the media and their own staff, because their meaning is at best vague and inaccurate and at worst deceptive.
The same translators call potential risk, very often used by the pharmaceutical industry, a subtle euphemism and so hard to detect: the adjective softens the clearly negative meaning of the noun, as it does in potential threat or potential danger. Risk, like threat, already contains the idea of possibility or probability, so potential is redundant: something entails risk or it does not. Something can be a risk whatever the outcome of one occasion: someone driving while drunk may usually arrive safe and sound, but that does not mean drink-driving entails no risk. If sudden death has been associated with a drug in some patients, the risk exists and may affect everyone who receives the drug, however low its incidence, so a potential risk of sudden death only softens the phrase, and such a lack of accuracy leads to deceptive information and deceptive advertising. This may be a risk and it may be harmful are redundant and euphemistic in the same way, as is in general any statement that combines may with risk or threat, and any statement that softens its main noun with an adjective is a euphemism; there may be a potential risk of… is redundant three times over. The translators advise omitting an unnecessary may, so that the drug may cause skin rash in about 34% of patients is translated as a statement that the drug causes skin rash in about 34% of patients, and they hold that authors and translators should strive to avoid such imprecise terminology when writing English.
Terms used in place of a thought
Some words and phrases that are misused or used vaguely lead not just to unclear writing but, more importantly, to unclear thinking, in the experience of the psychologist Scott O. Lilienfeld as a writer and editor. His list of fifty psychological and psychiatric terms to avoid, published with colleagues, gives terms that its authors think should either be used more precisely or avoided, among them:
- a gene for a personality trait or psychiatric disorder, which is endemic in the popular media and, in Lilienfeld's view, misleading, because it may lead readers to think that major genes affect such complex behaviours, which evidence from genome-wide association studies increasingly suggests is very unlikely
- comorbidity, which can confuse because authors use it to mean covariation between two diagnoses in a sample or population or co-occurrence of two diagnoses in one person
- influence of or effect of gender, social class or education, causal language that often is not appropriate, depending on the study design
- love molecule for oxytocin, a description its authors call woefully simplistic
- no difference between groups for a difference that was not statistically significant, since failing to reject the null hypothesis does not confirm it; the article suggests considering no significant difference between groups or no significant correlation between variables instead
In Lilienfeld's view, vague terms such as medical model or scientific method can sometimes almost substitute for deeper thought or understanding, "like a punt", and in his own teaching he found that he might throw out a term he had not thought through deeply or clearly as a placeholder. The list is not a ban but a call to use these terms carefully or only in certain contexts: these authors adopt a medical model in the sense that followed by what is meant is fine, while these authors used a medical model can imply that disorders such as schizophrenia or depression are categorical rather than dimensional, that they have biological causes, or that they should be treated biologically or by a physician, and sometimes other things. In one editor's experience, digging into a text shows that its authors think they have said one thing when they have not ruled out a hundred other possibilities. Lilienfeld holds that a field such as psychology, which deals with fuzzy concepts that lend themselves to misunderstanding, has more reason, not less, to be very clear about what it does and does not mean. Every writer will be a little imprecise from time to time, which Lilienfeld considers fine, but the goal is to be as clear and careful in communication as possible to prevent public misunderstandings.
Gold standard, common in the eating disorders field, is on both the fifty-term list and a list of thirteen terms often used in problematic ways in that field: even well-validated research instruments and treatments are imperfect, so authors should avoid the shorthand and, as applicable, describe an instrument as extensively validated or a treatment as demonstrably superior to other treatments.
Phrases that gesture at a gap or at novelty
Credibility is key to building and maintaining trust in science, and the phrases a writer uses can strengthen or tarnish it. Medical writers support the credibility of science by basing their logic on previous findings, describing methods and results clearly and keeping their interpretations apart from the data. Vague language about their logic and interpretations can undermine the content and tarnish the credibility of both the writing and the work, and of its many forms, three phrases stand out, in one writer's view, as the commonest that tarnish credibility: "little is known", "to our knowledge" and "first to show".
"Little is known" fails in three ways. Most topics have had many papers published on them, and the question, gap or problem a work addresses is bounded by what is known, so the phrase does a poor job of defining it and can lead readers to think that the writer does not know the literature well or is overstating the case; it is often preceded or followed by a fuller account of what is known, which contradicts it and tarnishes credibility; and it leaves the reader asking how little: one paper, a dozen that miss the specific question, or an extensive literature whose findings the work aimed to confirm a new tool will validate. A convincing argument defines instead the specific question, gap or problem the work aimed to address. For example, "Little is known about how the drug reduces cell proliferation in lung cancer" becomes "Although we know that the drug reduces proliferation of A549 cells, we do not know the mechanism by which the gene regulates this process".
"To our knowledge" implies that others may know what the writer does not, which can cast doubt on whether the writer knows the field, and it suggests an unwillingness to be held accountable if the information becomes known, a lack of responsibility that can damage the writer's reputation and the work's credibility. The phrase is better omitted, and only known information shared. An example rewrites "To our knowledge, researchers have not uncovered how the drug reduces cell proliferation of lung cancer cells" as "Although we do not fully understand how the drug reduces cell proliferation in lung cancer, researchers have linked the drug to cell mitosis in other cancer cells."
Some writers believe "first to show" persuades by highlighting the novelty or significance of the work, but it suggests knowledge of everything in the field, published or unpublished, which is hard to guarantee, and it can make the writer seem overconfident or to be overstating the case. Unless the writer is indisputably certain of being first, the phrase is better avoided. An example rewrites "We are the first to show that activation of the gene increases proliferation of A549 cells" as "We showed that activation of the gene increases proliferation of A549 cells."
Writers often use all three in one document, building the case with "little is known", protecting their reputation with "to our knowledge" and claiming novelty with "first to show", and together the phrases can have a cumulative effect that erodes credibility in the writing and in the work, so all three are better avoided, alone or together. A paired example rewrites "Little is known about how the drug affects plaque formation associated with Alzheimer's disease…. To our knowledge, this study is the first to show that the drug slows plaque formation" as "Although the drug reduces beta-amyloid, we do not know if the drug slows plaque formation associated with Alzheimer's disease…. In this study, we showed that the drug slows plaque formation".
Words with a legal consequence: obvious
In US patent law, non-obviousness is one qualification for the grant of a patent, so the word obvious in a research notebook, or worse in a formal publication, could easily make a patent grant invalid and might even devalue years of research and millions of dollars of investment. Researchers at one pharmaceutical company were taught patent law and how to write research notebooks in anticipation of potentially patentable discoveries, and one rule was never to use obviously, or any other word portraying a concept as not novel. One editor who worked there as a researcher deletes obvious scrupulously when editing and replaces obviously with words such as surprisingly, unexpectedly, contrary to our hypothesis and interestingly.
In that editor's view, whether most science writers use such replacements out of awareness of the legal implications of obvious or out of copying the lingo they have read, it is materially important to qualify a result as surprisingly non-obvious. One small word can have a regulatory impact.
Words that date a claim: recently and new
Recently leaves open how long a new drug is still considered recently approved: six months, a year or some other time frame. One usage editor holds that vagueness is never a virtue in medical writing, and that new is as ambiguous and open to interpretation as recently. Without knowing the context, she would give a more informative indication of the time involved, even if only "within the past year".
The writer should keep in mind the forthcoming date of publication rather than the date of the manuscript, which takes a little homework; where a drug has been approved by the FDA, the approval date is readily ascertainable from the agency itself or from the National Library of Medicine or its websites.